Research: Wireless Communications

This project introduces a novel wireless network architecture that enables intelligent, spectrum and energy-efficient communications in dynamic RF environments. By integrating spatio-temporal spectrum sensing, dynamic spectrum access (DSA), and neuromorphic computing, mobile users can perform short-range, local communications with minimal energy consumption. The research develops DSA-enabled heterogeneous networks (HetNets), cooperative resource allocation techniques, and neuromorphic hardware to address high computational complexity with ultra-low power. A hardware-software testbed will validate the system. The project also includes educational initiatives on energy-efficient communications and neuromorphic design, fostering industry outreach and broadening participation in next-generation wireless technologies.

This project introduces sensing-based device-to-device (D2D) communication to address the rapid increase in mobile data traffic. In this approach, user devices perform spatial spectrum sensing to identify temporal and spatial transmission opportunities within cellular network bands. By using non-occupied cellular spectrum while protecting existing cellular users, sensing-based D2D can greatly enhance network spectral efficiency. The project develops a theoretical and practical framework, explores optimal system design and resource allocation, and evaluates performance through software and hardware testbeds.

This project develops the fundamental research needed to integrate terrestrial and non-terrestrial networks into unified Ground and Air Integrated Networks (GAINs), addressing the limits of current mobile broadband coverage. By designing waveforms in the delay-Doppler domain, the research enables machine learning to operate at NextG speeds. The project includes waveform design, multi-agent reinforcement learning for resilient scheduling, and distributed, resilient computing tailored to heterogeneous GAIN environments. Software and hardware testbeds will demonstrate the concepts. This interdisciplinary effort aims to create intelligent, robust communication and computing technologies for future mobile broadband networks.

This project aims to advance the use of artificial intelligence and machine learning in wireless communication by combining data-driven learning methods with conventional model-based approaches. Focusing on physical-layer problems, the research explores how structural knowledge inherent in wireless networks can guide the design of more efficient and reliable learning algorithms. Three key areas are investigated: symbol detection in MIMO interference channels, massive MIMO with low-resolution ADCs, and MIMO-OFDM waveform design under nonlinear RF effects. The project seeks to develop tailored algorithms that offer improved performance and stronger theoretical guarantees.

This project develops the Intelligent Dynamic spEctrum Access (IDEA) framework to significantly improve spectrum utilization, energy efficiency, and coexistence among primary, secondary, and passive wireless systems. IDEA integrates neuromorphic spiking neural network hardware for ultra-low-power on-device intelligence, deep reinforcement learning for dynamic access, and advanced spectrum sensing for fast emitter detection. The framework supports heterogeneous radios while protecting passive services essential for scientific research. Through software and hardware testbeds, the project enables efficient, agile spectrum access and holistic system optimization, benefiting next-generation wireless networks and high-demand applications such as IoT, smart cities, and autonomous systems.